Journal of neuroimaging : official journal of the American Society of Neuroimaging · 2025
T1-Dark Рим как знак новых и хронически активных множественных склерозов: Серийное исследование с частым 7T MRI.
СтатьяНазвание и аннотация: Машинный перевод · не проверен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Marshall Madeleine, Aphiwatthanasumet Kingkarn, Mougin Olivier, Stadelmann Christine, Morgan Paul S, Dineen Rob A, Gowland Penny, Evangelou Nikos, Clarke Margareta A
Аннотация
ЦЕЛЬ И СПРАВОЧНАЯ ИНФОРМАЦИЯ: Хронические активные множественные повреждения склероза (MS) представляют собой особо разрушительную группу повреждений патологии. Однако их изображения коррелируют, в том числе с повреждениями парамагнитных ободов (ПБД), которые обнаруживаются на изображениях, взвешенных по степени чувствительности, и с трудом поддаются применению в клинической практике. В ходе этого исследования в продольном разрезе изучается распространенность и временная динамика нового маркера изображений, T1-dark В области количественного картирования уязвимости (QSM) были отмечены ободья и их связь с PRL. МЕТОДЫ: Four нелечимые лица, имеющие MS 7-Tesla Сканирование МРТ six в течение периода времени 36 недель. Были проанализированы новые и ранее существовавшие повреждения с точки зрения присутствия и временнóй эволюции T1-dark и ободья QSM. Количественные значения Т1 были получены с использованием карт В1, и была оценена взаимосвязь между состоянием обода и размером повреждения. РЕЗУЛЬТАТЫ: 159 - фоновые повреждения, 22 (В тыс.14%) имели T1-dark ободья, 11 (В тыс.7%) имели ободья QSM, и five Повреждения были обе. T1-dark и ободья QSM продемонстрировали временные изменения, T1-dark обода перед появлением нового обода QSM в three Из four (В тыс.75%Повреждения. Eleven Из 20 (В тыс.55%) Недавно образовавшиеся повреждения T1-dark ободья с a T1-dark обод, присутствующий при всех новых повреждениях над уровнем моря 100 мм3. Небольшие новые повреждения не имеют видимых ободьев, однако их общие значения Т1 соответствуют значениям более крупных повреждений T1-dark Кривые, подразумевающие общие патологические процессы. ВЫВОДЫ: T1-dark Кольца были более распространенными, чем кольца QSM, при этом более часто встречаются недавно образовавшиеся повреждения. Мы предлагаем, чтобы они представляли отёк и воспаление, связанные с ранними стадиями хронического активного поражения, видимыми, несмотря на накопление железа, а не из-за этого.
Оригинал публикации
T1-Dark Rim as a Marker of New and Chronic Active Multiple Sclerosis Lesions: A Serial Study With Frequent 7T MRI.
BACKGROUND AND PURPOSE: Chronic active multiple sclerosis (MS) lesions represent a particularly destructive subset of lesions on pathology. However, their imaging correlates, including paramagnetic rim lesions (PRLs) detected on susceptibility-weighted imaging, lack sensitivity and are difficult to implement in clinical practice. This exploratory, longitudinal study investigates the prevalence and temporal dynamics of a novel imaging marker, T1-dark rims, and their relationship with PRLs observed on quantitative susceptibility mapping (QSM).
METHODS: Four untreated people with MS underwent 7-Tesla MRI scanning six times over a period of 36 weeks. New and pre-existing lesions were analyzed for the presence and temporal evolution of T1-dark and QSM rims. Quantitative T1 values were derived using B1 maps, and the relationship between rim status and lesion size was evaluated.
RESULTS: Of the 159 baseline lesions, 22 (14%) had T1-dark rims, 11 (7%) had QSM rims, and five lesions had both. T1-dark and QSM rims showed temporal changes, with T1-dark rims preceding new QSM rim appearance in three out of four (75%) lesions. Eleven out of 20 (55%) newly formed lesions had T1-dark rims, with a T1-dark rim present in all new lesions over 100 mm3. Small new lesions lacked discernible rims, but their overall T1 values aligned with those of larger lesion T1-dark rims implying shared pathological processes.
CONCLUSIONS: T1-dark rims were more common than QSM rims, with greater prevalence in newly formed lesions. We propose they represent edema and inflammation associated with early stages of chronic active lesion formation visible despite, not because of, iron accumulation.
Перевод полного текста: Перевод в очереди
Читать найденный полный текст
https://creativecommons.org/licenses/by/4.0/ This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
Introduction
Multiple sclerosis (MS) is an autoimmune disease characterized by the presence of demyelinating lesions in the CNS. Postmortem studies of acute MS lesions have revealed that oligodendrocyte apoptosis, immune cell influx, and inflammation occur at the initial stages of lesion formation around a central vein, accompanied by the opening of the blood–brain barrier (BBB) [ 1 , 2 ]. Consequent fluid leakage into the perivascular space leads to edema [ 3 ], followed by destruction and phagocytosis of myelin at the center of the lesion, which may continue at the lesion edges during the chronic active stage [ 1 ].
Being able to characterize specific pathological processes associated with different stages of MS lesion evolution in vivo is one of the main goals of neuroimaging research in MS [ 4 ]. MRI is routinely used to identify new lesions and monitor the evolution of existing lesions at the time of diagnosis and throughout disease duration. With the use of different MRI sequences, significant research progress has been made toward visualizing specific pathological processes beyond merely just visualizing the presence of lesions.
Analysis of dynamic contrast enhancement has proposed two patterns of gadolinium enhancement seen in newly formed lesions. Initially, center‐out (also called centrifugal) enhancement is detected, which is thought to reflect the outward formation of a lesion from a central vein. Within days, this shifts to the periphery‐in (also called centripetal) pattern [ 5 , 6 ]. This is thought to reflect the closure of the BBB around the central vein and opening of the BBB in the peripheral vasculature as the lesion expands, with a concomitant shift in inflammation from the center to the periphery. Once the BBB is closed, accompanied by the resolution of gadolinium enhancement on MRI, a proportion of lesions evolve into chronic active lesions (CALs).
CALs have persistent, active inflammation at the edge of the demyelinated lesion thought to be driven by iron‐laden macrophages and microglia [ 7 ]. As a result, some CALs can be identified on susceptibility‐weighted imaging (SWI) as paramagnetic rim lesions (PRLs), which are characterized by the presence of a hypointense or hyperintense rim (depending on the sequence used) [ 8 ]. Pathological studies report that up to 57% of lesions are chronic active [ 9 ], whereas PRLs represent around 10% of all lesions [ 10 ].
Recently, Naval‐Baudin et al. described a novel imaging sign on 3T high‐resolution turbo field echo T 1 ‐weighted imaging, termed the T 1 ‐dark rim [ 11 ]. The authors suggested that T 1 ‐dark rims have higher sensitivity for detecting PRLs compared to SWI, including phase maps, and that T 1 ‐weighted imaging may therefore offer a more sensitive, alternative method of imaging CALs.
Currently, the prevalence and temporal evolution of T 1 ‐dark rims remain unknown, and their relation to PRLs requires further validation. The present study considers a dataset obtained by frequent 7 Tesla (T) scanning of untreated patients with MS acquiring magnetization prepared 2 rapid gradient echo (MP2RAGE), a T 1 ‐weighted sequence that is more resistant to magnetic field inhomogeneities [ 12 ], and quantitative susceptibility mapping (QSM), which has better pathological sensitivity and specificity to iron than conventional SWI or phase maps [ 13 ]. Using this dataset, we have been able to perform an exploratory, longitudinal study reporting the prevalence and evolution of T 1 ‐dark rims and PRLs in nascent and established white matter lesions.
Methods
Participants and Image Acquisition
The study was granted approval by a local ethics committee (05/Q2404/44), and all subjects signed an informed consent form prior to study enrolment. We analyzed scans from an exploratory pilot study with frequent scanning of four untreated patients with MS considering disease‐modifying therapy [ 14 ].
Patients underwent MRI scanning six times over 36 weeks, except for one patient who had five visits. Imaging was performed on a 7T Philips Achieva scanner with a single‐transmit 32‐channel head coil (Nova Medical). The specific acquisition parameters for each sequence are listed in Table 1 . B 1 maps were used to correct data for inevitable field variations within the scanner at 7T. We did not administer gadolinium contrast agent in this study.
TABLE 1 MRI sequence parameters. Contrast Resolution (mm 3 ) FOV (mm) (AP, RL, SI) SENSE (RL, AP) Readout BW (Hz) TR (ms) TE (ms) TI (ms) FA (°) Duration (min) FLAIR 0.6 × 0.6 × 0.6 192 × 180 × 120 1.5, 1 1599.0 4800 256 1650 90 5.28 T 1 w 0.6 × 0.6 × 0.6 200 × 180 × 120 2.2, 2 115.2 13 6 787 8 11.58 T 2 *w, QSM 0.6 × 0.6 × 0.6 192 × 185 × 70 2.5, 2 506.4 50 20 14 6.77 B 1 map 3.2 × 4.0 × 4.0 205 × 180 × 132 No 4271.7 20, 120 5 60 3.05 Abbreviations: AP, anterior–posterior; BW, bandwidth; FA, flip angle; FLAIR, fluid‐attenuated inversion recovery; FOV, field of view; QSM, quantitative susceptibility mapping; RL, right–left; SENSE, sensitivity encoding; SI, superior–inferior; TE, echo time; TI, inversion time; TR, repetition time; w, weighted. John Wiley & Sons, Ltd.
MRI sequence parameters.
Abbreviations: AP, anterior–posterior; BW, bandwidth; FA, flip angle; FLAIR, fluid‐attenuated inversion recovery; FOV, field of view; QSM, quantitative susceptibility mapping; RL, right–left; SENSE, sensitivity encoding; SI, superior–inferior; TE, echo time; TI, inversion time; TR, repetition time; w, weighted.
Image Postprocessing
MP2RAGE images were created using an in‐house‐developed approach [ 15 ]. T 2 *‐weighted fast field echo QSM data were reconstructed using STI Suite (v3.0, Berkeley, California, https://people.eecs.berkeley.edu/~chunlei.liu/software.html ) [ 16 ]. A brain mask was created using the Brain Extraction Tool in FSL (v6.0.7.6., Oxford, UK, https://fsl.fmrib.ox.ac.uk/fsl/docs/#/ ) [ 17 ]. We used the Laplacian‐based phase unwrapping to process the signal phase, the V‐SHARP method with variable‐radius kernel at five voxels to reduce the sophisticated harmonic artifact for phase data, and the improved sparse linear equation and least‐square algorithm to generate susceptibility maps for each subject. QSM values were calculated relative to cerebrospinal fluid (CSF). This was achieved by averaging QSM values from four regions of interest in the CSF and subtracting that value from values found in other regions of interest. All images were rigidly registered to the baseline MP2RAGE scan using FSL FLIRT (Oxford, UK, https://fsl.fmrib.ox.ac.uk/fsl/docs/#/ ).
Rim Lesion Definitions and Analysis
T 1 ‐dark rims were identified on MP2RAGE images according to the criteria used by Naval‐Baudin et al., which states that a hypointense ring of signal surrounding the lesion must be visible on two or more consecutive slices or on two orthogonal planes [ 8 ]. Any rims that did not fulfill the criteria (e.g., T 1 ‐dark rim was clearly visible but only present on a single slice) were also recorded.
Paramagnetic rims manifesting as hyperintense rims on QSM were identified in line with the consensus definition [ 8 ]. A lesion with a paramagnetic rim can only be called a PRL if it does not enhance on a postcontrast scan or, in the absence of gadolinium contrast administration (as was the case in our study), the lesion's presence must be verified on a scan performed 3 months prior. To simplify terminology, we use the term “lesion with a QSM rim” and specify the number of possible PRLs (any lesions with a QSM rim) and true PRLs (lesions verified to be at least 3 months old with a QSM rim).
All the scans were reviewed by a single rater (M.M.) and subsequently reviewed by a senior researcher with 7 years of experience in PRL identification (M.A.C.). Any disagreements were discussed, and a consensus agreement was reached regarding the rim status of the lesion.
Cross‐Sectional Analysis of Pre‐Existing Lesions
Baseline white matter lesions visible on fluid‐attenuated inversion recovery (FLAIR) images were manually segmented, and their volumes were recorded using ITK‐SNAP (v.4.0.2, Philadelphia, PA, http://www.itksnap.org/pmwiki/pmwiki.php ). The presence of a T 1 ‐dark rim and/or QSM rim was recorded for each lesion.
Longitudinal Analysis of Pre‐Existing Lesions
The pre‐existing baseline lesions were reviewed at the final study visit to determine whether a T 1 ‐dark rim and/or QSM rim were present. Only lesions with a QSM rim on the follow‐up scan (irrespective of QSM rim presence at baseline) were classed as true PRLs, in line with the consensus definition. Lesions with a QSM rim on the baseline scan only were classed as possible PRLs.
Longitudinal Analysis of New Lesions
Each scan was reviewed to identify new lesions that appeared during the study follow‐up (visits 2–6). These were segmented on both FLAIR and MP2RAGE images, and the volume of each lesion was calculated. All new lesions were then assessed for the presence of T 1 ‐dark and QSM rims on all subsequent follow‐up scans. Only lesions with a QSM rim visible across three or more visits were classed as true PRLs.
Additionally, the detection of the T 1 ‐dark rim was compared between the MP2RAGE and B 1 maps to determine whether there was agreement between the two sequences. Regions with long T 1 values appear hypointense on the MP2RAGE images and hyperintense on B 1 maps.
Lastly, B 1 maps were used to quantify longitudinal changes in T 1 values in each new lesion. For all new lesions with a T 1 ‐dark rim, the T 1 ‐dark rim and lesion core were manually segmented. The masks were then applied to B 1 maps, and mean T 1 values were calculated for (1) new rimless lesions (based on the whole lesion mask), (2) the rim only of the lesions with a T 1 ‐dark rim, and (3) the lesion core only of the lesions with a T 1 ‐dark rim. This was done across all subsequent time points. Comparisons of the T 1 values between the rimless and rim‐positive lesions were carried out within the same patients to account for possible intersubject and interscan differences.
Statistical Analysis
Considering the exploratory nature of the study and the small sample size, no formal statistical analyses were conducted. Descriptive statistics are provided.
Results
The baseline characteristics of the patients included in this study are presented in Table 2 . The average interval between visits was 7.6 weeks, and the average total follow‐up period was 36 weeks.
TABLE 2 Patient characteristics. Patient Sex Age Disease subtype Time since disease onset (months) Number of lesions Number of new lesions during the follow‐up 1 M 23 CIS 5 56 9 2 F 39 RRMS 261 22 1 3 F 35 CIS 1 32 2 4 F 27 RRMS 25 49 8 Abbreviations: CIS, clinically isolated syndrome; F, female; M, male; RRMS, relapsing‐remitting multiple sclerosis. John Wiley & Sons, Ltd.
Patient characteristics.
Abbreviations: CIS, clinically isolated syndrome; F, female; M, male; RRMS, relapsing‐remitting multiple sclerosis.
Cross‐Sectional Analysis
In total, 159 pre‐existing white matter lesions were identified on the baseline scans. A patient‐wise breakdown of all rim‐positive lesions is provided in Table 3 . Examples of T 1 ‐dark and QSM rims are shown in Figure 1 .
TABLE 3 A summary of all rim‐positive lesions recorded at baseline and follow‐up. Number of rim‐positive lesions Baseline Final follow‐up (36 weeks later) Patient T 1 ‐dark QSM T 1 ‐dark QSM 1 19 8 14 7 2 1 0 1 0 3 2 2 2 2 4 0 1 2 2 Total lesions (%) 22 (13.8%) 11 (6.9%) 19 (11.9%) 11 (6.9%) Abbreviation: QSM, quantitative susceptibility mapping. John Wiley & Sons, Ltd.
A summary of all rim‐positive lesions recorded at baseline and follow‐up.
Abbreviation: QSM, quantitative susceptibility mapping.
Lesions with T 1 ‐dark rims on magnetization prepared 2 rapid gradient‐echo (marked by orange arrows) (left) and the corresponding quantitative susceptibility mapping (QSM) image showing a single paramagnetic rim lesion (marked by purple arrow) with a clear vein passing through its center (right). A magnification of the QSM‐rim is provided in the top left corner.
T 1 ‐Dark Rims Are More Common Than QSM Rims
A total of 22 lesions (13.8%) were classed as having a T 1 ‐dark rim and 11 (6.9%) as having a QSM rim. A further six lesions had a clear T 1 ‐dark rim on one slice only and so were classified as rimless according to our protocol.
A total of 17 lesions had a T 1 ‐dark rim only, six had a QSM rim only, and five lesions had both types of rims. Compared to the signal intensity of nearby white matter lesions, five of the six lesions with a QSM rim only appeared globally very hypointense on the MP2RAGE scans and one had patches of hypointense signal.
Lesions With Rims Are Larger Than Rimless Lesions
Figure 2 shows the volumes of all lesions on the baseline scans according to rim status (T 1 ‐dark, QSM, or both). The largest lesions had both T 1 ‐dark and QSM rims, while most of the rimless lesions were noted to have a volume under 100 mm 3 .
Baseline (pre‐existing) lesions with a T 1 ‐dark rim only ( n = 17), quantitative susceptibility mapping (QSM) rim only ( n = 6), or both T 1 ‐dark and QSM rims ( n = 5) had a larger fluid‐attenuated inversion recovery (FLAIR) volume than lesions without either rims ( n = 131). n , number of lesions.
Longitudinal Analysis of Pre‐Existing Lesions
A comparison of the total and patient‐wise numbers of pre‐existing lesions with T 1 ‐dark and QSM rims over a short‐term follow‐up of 36 weeks is shown in Table 3 .
Lesions With T 1 ‐Dark and QSM Rims Show Dynamic Changes on Short‐Term Follow‐Up
On the final follow‐up scan, a total of five pre‐existing lesions had both a T 1 ‐dark and QSM rim. Two lesions had T 1 ‐dark and QSM rims persisting from baseline to follow‐up. Three other lesions had a T 1 ‐dark rim at baseline, but the QSM rim only appeared on the follow‐up scan. Four other T 1 ‐dark rims developed during the study without QSM rims. At baseline, one of these appeared markedly hypointense on the MP2RAGE scan. Table 4 summarizes the temporal evolution of all pre‐existing lesions with a T 1‐ dark or QSM rim at the first and/or last study visit. Out of all the pre‐existing lesions studied, four were possible PRLs and 11 were true PRLs.
TABLE 4 Summary of the longitudinal evolution of lesions with T 1 ‐dark rims and paramagnetic rim lesions between baseline and last follow‐up. Rim lesion type Number of rims persisting Number of rims appearing Number of rims disappearing T 1 ‐dark 14 5 8 QSM 7 4 4 Abbreviation: QSM, quantitative susceptibility mapping. John Wiley & Sons, Ltd.
Summary of the longitudinal evolution of lesions with T 1 ‐dark rims and paramagnetic rim lesions between baseline and last follow‐up.
Abbreviation: QSM, quantitative susceptibility mapping.
T 1 ‐Dark Rim Lesions Can Precede the Appearance of QSM Rims
Four pre‐existing lesions had a QSM rim appear at the final visit. Three of those QSM rims (75%) were preceded by a T 1 ‐dark rim at baseline. The fourth lesion presented patches of hypointense signal in the lesion core on MP2RAGE at baseline. One baseline lesion with a QSM rim had a marked T 1 hypointense signal throughout the lesion core and developed a T 1 ‐dark rim at the follow‐up visit.
T 1 ‐Dark Rims Persist Longer Than QSM Rims
In two of the five (40%) pre‐existing lesions with both a T 1 ‐dark rim and a QSM rim at baseline, the rims persisted 36 weeks later. For the remaining three lesions (60%), the T 1 ‐dark rim persisted for the whole study period, but the QSM rim disappeared.
Longitudinal Analysis of New Lesions
A total of 20 new lesions appeared over the course of the study. Figure 3 describes the T 1 ‐dark and QSM rim status of each new lesion.
The temporal evolution of new lesions with T 1 ‐dark and quantitative susceptibility mapping (QSM) rims. The lesions are arranged according to volume (from smallest to largest). Largest lesions had both T 1 ‐dark and QSM rims, while smaller lesions had a transient T 1 ‐dark rim or were markedly hypointense on the magnetization prepared 2 rapid gradient‐echo image. No new lesions had a QSM rim without also having a T 1 ‐dark rim. Lesions not analyzable on QSM across the whole study duration were excluded. Two lesions were not analyzable at Visit 5 due to poor image quality. Lesions without data for visit 6 are all from Patient 4, who only had five visits.
Most New Lesions Have a T 1 ‐Dark Rim
Eleven lesions met the criteria for a T 1 ‐dark rim, while a further five had a T 1 ‐dark rim on one slice only and were classified as rimless. In eight of the 11 nascent lesions with a T 1 ‐dark rim, the rim appeared at the first appearance of the lesion (see Figure 3 ). Five of the T 1 ‐dark rims persisted for at least three visits and were present at the end of the data collection period. Another four T 1 ‐dark rims disappeared after a single visit, while two lesions with T 1 ‐dark rims only appeared on the final visit and consequently were not able to be followed up.
New PRLs Are Accompanied by a T 1 ‐Dark Rim
Five new lesions (25%) had a rim on QSM (see Figure 3 ). Four of these QSM rims appeared simultaneously to a T 1 ‐dark rim, and the fifth QSM rim appeared one visit after the appearance of a T 1 ‐dark rim. No new lesions had a QSM rim without also having a T 1 ‐dark rim. Of the five new lesions with QSM and T 1 rims, four had both rims still present at the final visit, while one lesion showed the QSM rim disappearing at the final visit with the T 1 rim still present. Two of the five lesions with QSM rims had a rim visible across three visits and were therefore classed as true PRLs.
The Presence of Rims Is Associated With the Size of the New Lesion
The presence of rims was related to the size of the lesions, as measured on FLAIR and MP2RAGE scans (see Figure 4 ). The largest lesions had both T 1 ‐dark and QSM rims (see Figure 4A,C ). Medium‐sized lesions had T 1 ‐dark rims only. The smallest lesions were rimless. The persistence of T 1 ‐dark rims was also related to the size of the lesion (see Figure 4B,D ).
Magnetization prepared 2 rapid gradient‐echo (MP2RAGE) (top row) and fluid‐attenuated inversion recovery (FLAIR) (bottom row) volumes of new lesions according to rim status. (A) Average MP2RAGE volume of new rimless, T 1 ‐dark rim only, and T 1 ‐dark and quantitative susceptibility mapping (QSM) rim lesions. Three lesions without a T 1 rim were unable to be assessed on QSM and were therefore excluded. (B) Average MP2RAGE volume of new lesions according to the temporal evolution of the T 1 ‐dark rim: rimless, temporary T 1 ‐dark rim, or persistent T 1 ‐dark rim. Two lesions with rim appearing at the final visit were excluded. (C) Average FLAIR volume of new rimless, T 1 ‐dark rim only, and T 1 ‐dark and QSM rim lesions. (D) Average FLAIR volume of new lesions according to the temporal evolution of the T 1 ‐dark rim: rimless, temporary T 1 ‐dark rim, or persistent T 1 ‐dark rim.
MP2RAGE and B1 Maps Show High Agreement in Detecting T 1 ‐Dark Rims
T 1 ‐dark rims had consistently longer T 1 values than the lesion core (data not shown). New lesions on B 1 maps corresponded with what was observed on MP2RAGE in 18 of the 20 lesions (90%). One lesion only showed a rim on one slice on the MP2RAGE (and hence had been classed as rimless) but showed a rim on two slices on the B 1 map. The other discrepancy was a large edematous lesion, which was classed as having a T 1 ‐dark rim on MP2RAGE but was globally hyperintense on B 1 .
The Core of New Rimless Lesions Is as Hypointense as the Rim of T 1 ‐Dark Rim Lesions
All the new lesions without a T 1 ‐dark rim were small with a volume of less than 100 mm 3 (see Figure 4 ). On MP2RAGE images, their core appeared diffusely hypointense with signal intensity similar to the rim rather than the core of the large, T 1 ‐dark rim lesions. Calculations of T 1 signal from B 1 maps confirmed these findings. Moreover, new rimless lesions appeared to have higher T 1 values when compared to similar‐sized, pre‐existing, rimless lesions. Example values from Patient 1, who had at least three lesions of each type, are shown in Figure 5 . Consistent with pre‐existing small lesions having reduced T 1 values, the T 1 values of the small new lesions decreased over time (data not shown).
T 1 values of new rimless lesions (whole lesion), lesion core and rim of new T 1 ‐dark rim lesions (provided separately), and pre‐existing rimless lesions (whole lesion) for Patient 1. X shows mean values.
Discussion
This 7T MRI study presents novel findings on the prevalence and short‐term evolution of a newly described imaging marker, T 1 ‐dark rims, and their relationship with PRLs detected on QSM. Our findings suggest that T 1 ‐dark rims are sensitive at detecting, what we believe are, active MS lesions (either acute or chronic active).
We confirm observations from the first study to describe T 1 ‐dark rims that they are more numerous than PRLs [ 11 ]. Using 7T imaging and QSM, we detected twice as many T 1 ‐dark rims compared to PRLs, whereas in the 3T Naval‐Baudin et al. clinical study, T 1 ‐dark rim lesions outnumbered PRLs by around a quarter. In addition to the MRI field strength, our increased T 1 ‐dark rim detection rate is likely due to the differences in our methodologies; Naval‐Baudin and colleagues used SWI and phase imaging, which are not as pathologically specific to iron and, particularly phase, can lead to false positive findings [ 10 ]. In addition, by analyzing all brain lesions in a few patients with early MS, we were able to report the previously unknown prevalence of T 1 ‐dark rims (14% of all lesions).
Our investigation of new lesions revealed a striking prevalence of T 1 ‐dark rims. Over half of the new lesions displayed a T 1 ‐dark rim, a proportion that rose to 80% when considering lesions exhibiting a rim on at least one slice. Our quantitative analyses confirmed visual observations that T 1 ‐dark rims have longer T 1 values than their lesion core. Several pathological processes associated with new lesion development can increase T 1 signal, including demyelination, edema, phagocyte invasion, and hypercellularity [ 4 , 18 ]. Based on the location of the signal, at the periphery of new lesions, we postulate that marked edema associated with acute inflammation, rather than demyelination, is the primary driver of the T 1 ‐dark rim signal [ 19 ].
Notably, the remaining new lesions without a T 1 ‐dark rim had small volumes, all under 100 mm 3 . Small lesions may lack visible rims due to their size (partial volume artifacts), and this could be tested by imaging at higher spatial resolution. However, comparing the T 1 values of these small, new, rimless lesions to the rims and cores of larger T 1 ‐dark rim‐positive lesions confirmed that their core T 1 values matched the values observed at the T 1 ‐dark rim, suggesting a shared pathological origin. Furthermore, small, new, rimless lesions had higher T 1 values than size‐matched, rimless, older lesions, possibly indicating decreased inflammation and edema, consistent with pathological findings of lesions in the chronic inactive stage [ 20 ]. A subsequent T 1 reduction over the follow‐up period likely reflects this progression.
T 1 ‐dark rims were either transient (visible at only one visit) or persistent. A quarter of new lesions had transient T 1 ‐dark rims, suggestive of subsiding edema. These lesions never developed a paramagnetic rim, suggesting a less aggressive disease pathology. Persistent T 1 ‐dark rim lesions were always larger and accompanied by a paramagnetic rim, appearing either concurrently or one visit later. This indicates the presence of ongoing inflammation and tissue destruction, suggesting T 1 ‐weighted imaging may offer a sensitive alternative for CAL detection, particularly as SWI is not currently recommended for MS monitoring.
While paramagnetic rims reflect iron‐laden macrophages and microglia, iron decreases T 1 signal [ 21 ], opposing the increased T 1 signal observed in T 1 ‐dark rims. Thus, the T 1 ‐dark rim is visible despite, not because of, iron accumulation. We speculate that paramagnetic rims may mark a more intense inflammatory and destructive process involving demyelination and macrophage‐mediated myelin (and iron) clearance, making them less common than inflammation alone at acute lesion borders. In another paper by Naval‐Baudin et al., the authors showed that deeply hypointense T 1 voxels, often in the lesion periphery, were almost exclusively present in PRLs compared to non‐PRLs [ 22 ]. The association of both T 1 ‐dark and paramagnetic rims with larger lesion sizes supports the hypothesis that both rim types are associated with a more aggressive disease pathology. These lesions may also represent a subset of slowly expanding lesions, a possibility requiring further investigation with longer follow‐up studies.
Future pathology and MRI studies will establish the role and long‐term evolution of T 1 ‐dark rims in MS. For now, it is interesting to speculate whether the high incidence of T 1 ‐dark rims in new lesions could potentially be used to provide evidence of dissemination in time (DIT) in patients presenting with suspected MS. Future studies could determine whether the simultaneous presence of rim‐positive and rim‐negative lesions can be used as evidence of DIT, similarly to the way gadolinium‐enhancing lesions are currently used. Moreover, T 1 ‐dark rims could have a role in the monitoring of lesions or disease progression. Particularly, T 1 ‐dark rims could act as surrogate markers of early and chronic inflammation and could therefore be used to assess the efficacy of novel disease‐modifying therapies such as Bruton's tyrosine kinase inhibitors, which act on both sides of the BBB. Given the existing, widespread use of T 1 ‐weighted imaging in routine MS monitoring and the detection of T 1 ‐dark rims by clinical 3T scanners [ 11 ], this new radiological sign could easily be implemented in clinical practice.
This analysis faces several limitations. First and foremost, our exploratory study included only four patients, and our results need to be validated in a larger sample size. Importantly, these data did not include contrast‐enhanced imaging, so it was not possible to definitively identify the acute stage of nascent lesions or determine a temporal correlation between T 1 ‐dark rims and the resolution of gadolinium enhancement. However, it is known that MS lesions rarely enhance beyond 2–3 months [ 23 ], and T 1 ‐dark rims were seen to persist for 5 months, until the end of the data collection period.
Furthermore, since nascent lesions with a persistent T 1 ‐dark rim and paramagnetic rim were still observed at the end of the study, interpretation of the long‐term evolution of T 1 ‐dark rims and the temporal correlation with paramagnetic rims is limited. Further longitudinal follow‐up and inclusion of imaging techniques such as proton density, to confirm whether T 1 ‐dark rims represent edema, and diffusion‐weighted imaging, to infer the underlying tissue integrity, will be useful in discerning the underlying pathology of T 1 ‐dark rims and differences between them and PRLs.
This study further characterized the novel imaging finding of T 1 ‐dark rims that surround some MS lesions. We propose that T 1 ‐dark rims may represent a marker of early lesions with ongoing edema/inflammation at the edges surrounding demyelination. Considering that T 1 ‐weighted imaging is one of the most commonly utilized sequences, T 1 ‐dark rims could easily be incorporated into routine, clinical assessment of patients with MS. However, future large, longitudinal studies are needed to further establish their role, evolution, and underlying pathology.
Conflicts of Interest
The authors declare no conflicts of interest
Multiple sclerosis (MS) is an autoimmune disease characterized by the presence of demyelinating lesions in the CNS. Postmortem studies of acute MS lesions have revealed that oligodendrocyte apoptosis, immune cell influx, and inflammation occur at the initial stages of lesion formation around a central vein, accompanied by the opening of the blood–brain barrier (BBB) [ 1 , 2 ]. Consequent fluid leakage into the perivascular space leads to edema [ 3 ], followed by destruction and phagocytosis of myelin at the center of the lesion, which may continue at the lesion edges during the chronic active stage [ 1 ].
Being able to characterize specific pathological processes associated with different stages of MS lesion evolution in vivo is one of the main goals of neuroimaging research in MS [ 4 ]. MRI is routinely used to identify new lesions and monitor the evolution of existing lesions at the time of diagnosis and throughout disease duration. With the use of different MRI sequences, significant research progress has been made toward visualizing specific pathological processes beyond merely just visualizing the presence of lesions.
Analysis of dynamic contrast enhancement has proposed two patterns of gadolinium enhancement seen in newly formed lesions. Initially, center‐out (also called centrifugal) enhancement is detected, which is thought to reflect the outward formation of a lesion from a central vein. Within days, this shifts to the periphery‐in (also called centripetal) pattern [ 5 , 6 ]. This is thought to reflect the closure of the BBB around the central vein and opening of the BBB in the peripheral vasculature as the lesion expands, with a concomitant shift in inflammation from the center to the periphery. Once the BBB is closed, accompanied by the resolution of gadolinium enhancement on MRI, a proportion of lesions evolve into chronic active lesions (CALs).
CALs have persistent, active inflammation at the edge of the demyelinated lesion thought to be driven by iron‐laden macrophages and microglia [ 7 ]. As a result, some CALs can be identified on susceptibility‐weighted imaging (SWI) as paramagnetic rim lesions (PRLs), which are characterized by the presence of a hypointense or hyperintense rim (depending on the sequence used) [ 8 ]. Pathological studies report that up to 57% of lesions are chronic active [ 9 ], whereas PRLs represent around 10% of all lesions [ 10 ].
Recently, Naval‐Baudin et al. described a novel imaging sign on 3T high‐resolution turbo field echo T 1 ‐weighted imaging, termed the T 1 ‐dark rim [ 11 ]. The authors suggested that T 1 ‐dark rims have higher sensitivity for detecting PRLs compared to SWI, including phase maps, and that T 1 ‐weighted imaging may therefore offer a more sensitive, alternative method of imaging CALs.
Currently, the prevalence and temporal evolution of T 1 ‐dark rims remain unknown, and their relation to PRLs requires further validation. The present study considers a dataset obtained by frequent 7 Tesla (T) scanning of untreated patients with MS acquiring magnetization prepared 2 rapid gradient echo (MP2RAGE), a T 1 ‐weighted sequence that is more resistant to magnetic field inhomogeneities [ 12 ], and quantitative susceptibility mapping (QSM), which has better pathological sensitivity and specificity to iron than conventional SWI or phase maps [ 13 ]. Using this dataset, we have been able to perform an exploratory, longitudinal study reporting the prevalence and evolution of T 1 ‐dark rims and PRLs in nascent and established white matter lesions.
Methods
Participants and Image Acquisition
The study was granted approval by a local ethics committee (05/Q2404/44), and all subjects signed an informed consent form prior to study enrolment. We analyzed scans from an exploratory pilot study with frequent scanning of four untreated patients with MS considering disease‐modifying therapy [ 14 ].
Patients underwent MRI scanning six times over 36 weeks, except for one patient who had five visits. Imaging was performed on a 7T Philips Achieva scanner with a single‐transmit 32‐channel head coil (Nova Medical). The specific acquisition parameters for each sequence are listed in Table 1 . B 1 maps were used to correct data for inevitable field variations within the scanner at 7T. We did not administer gadolinium contrast agent in this study.
TABLE 1 MRI sequence parameters. Contrast Resolution (mm 3 ) FOV (mm) (AP, RL, SI) SENSE (RL, AP) Readout BW (Hz) TR (ms) TE (ms) TI (ms) FA (°) Duration (min) FLAIR 0.6 × 0.6 × 0.6 192 × 180 × 120 1.5, 1 1599.0 4800 256 1650 90 5.28 T 1 w 0.6 × 0.6 × 0.6 200 × 180 × 120 2.2, 2 115.2 13 6 787 8 11.58 T 2 *w, QSM 0.6 × 0.6 × 0.6 192 × 185 × 70 2.5, 2 506.4 50 20 14 6.77 B 1 map 3.2 × 4.0 × 4.0 205 × 180 × 132 No 4271.7 20, 120 5 60 3.05 Abbreviations: AP, anterior–posterior; BW, bandwidth; FA, flip angle; FLAIR, fluid‐attenuated inversion recovery; FOV, field of view; QSM, quantitative susceptibility mapping; RL, right–left; SENSE, sensitivity encoding; SI, superior–inferior; TE, echo time; TI, inversion time; TR, repetition time; w, weighted. John Wiley & Sons, Ltd.
MRI sequence parameters.
Abbreviations: AP, anterior–posterior; BW, bandwidth; FA, flip angle; FLAIR, fluid‐attenuated inversion recovery; FOV, field of view; QSM, quantitative susceptibility mapping; RL, right–left; SENSE, sensitivity encoding; SI, superior–inferior; TE, echo time; TI, inversion time; TR, repetition time; w, weighted.
Image Postprocessing
MP2RAGE images were created using an in‐house‐developed approach [ 15 ]. T 2 *‐weighted fast field echo QSM data were reconstructed using STI Suite (v3.0, Berkeley, California, https://people.eecs.berkeley.edu/~chunlei.liu/software.html ) [ 16 ]. A brain mask was created using the Brain Extraction Tool in FSL (v6.0.7.6., Oxford, UK, https://fsl.fmrib.ox.ac.uk/fsl/docs/#/ ) [ 17 ]. We used the Laplacian‐based phase unwrapping to process the signal phase, the V‐SHARP method with variable‐radius kernel at five voxels to reduce the sophisticated harmonic artifact for phase data, and the improved sparse linear equation and least‐square algorithm to generate susceptibility maps for each subject. QSM values were calculated relative to cerebrospinal fluid (CSF). This was achieved by averaging QSM values from four regions of interest in the CSF and subtracting that value from values found in other regions of interest. All images were rigidly registered to the baseline MP2RAGE scan using FSL FLIRT (Oxford, UK, https://fsl.fmrib.ox.ac.uk/fsl/docs/#/ ).
Rim Lesion Definitions and Analysis
T 1 ‐dark rims were identified on MP2RAGE images according to the criteria used by Naval‐Baudin et al., which states that a hypointense ring of signal surrounding the lesion must be visible on two or more consecutive slices or on two orthogonal planes [ 8 ]. Any rims that did not fulfill the criteria (e.g., T 1 ‐dark rim was clearly visible but only present on a single slice) were also recorded.
Paramagnetic rims manifesting as hyperintense rims on QSM were identified in line with the consensus definition [ 8 ]. A lesion with a paramagnetic rim can only be called a PRL if it does not enhance on a postcontrast scan or, in the absence of gadolinium contrast administration (as was the case in our study), the lesion's presence must be verified on a scan performed 3 months prior. To simplify terminology, we use the term “lesion with a QSM rim” and specify the number of possible PRLs (any lesions with a QSM rim) and true PRLs (lesions verified to be at least 3 months old with a QSM rim).
All the scans were reviewed by a single rater (M.M.) and subsequently reviewed by a senior researcher with 7 years of experience in PRL identification (M.A.C.). Any disagreements were discussed, and a consensus agreement was reached regarding the rim status of the lesion.
Cross‐Sectional Analysis of Pre‐Existing Lesions
Baseline white matter lesions visible on fluid‐attenuated inversion recovery (FLAIR) images were manually segmented, and their volumes were recorded using ITK‐SNAP (v.4.0.2, Philadelphia, PA, http://www.itksnap.org/pmwiki/pmwiki.php ). The presence of a T 1 ‐dark rim and/or QSM rim was recorded for each lesion.
Longitudinal Analysis of Pre‐Existing Lesions
The pre‐existing baseline lesions were reviewed at the final study visit to determine whether a T 1 ‐dark rim and/or QSM rim were present. Only lesions with a QSM rim on the follow‐up scan (irrespective of QSM rim presence at baseline) were classed as true PRLs, in line with the consensus definition. Lesions with a QSM rim on the baseline scan only were classed as possible PRLs.
Longitudinal Analysis of New Lesions
Each scan was reviewed to identify new lesions that appeared during the study follow‐up (visits 2–6). These were segmented on both FLAIR and MP2RAGE images, and the volume of each lesion was calculated. All new lesions were then assessed for the presence of T 1 ‐dark and QSM rims on all subsequent follow‐up scans. Only lesions with a QSM rim visible across three or more visits were classed as true PRLs.
Additionally, the detection of the T 1 ‐dark rim was compared between the MP2RAGE and B 1 maps to determine whether there was agreement between the two sequences. Regions with long T 1 values appear hypointense on the MP2RAGE images and hyperintense on B 1 maps.
Lastly, B 1 maps were used to quantify longitudinal changes in T 1 values in each new lesion. For all new lesions with a T 1 ‐dark rim, the T 1 ‐dark rim and lesion core were manually segmented. The masks were then applied to B 1 maps, and mean T 1 values were calculated for (1) new rimless lesions (based on the whole lesion mask), (2) the rim only of the lesions with a T 1 ‐dark rim, and (3) the lesion core only of the lesions with a T 1 ‐dark rim. This was done across all subsequent time points. Comparisons of the T 1 values between the rimless and rim‐positive lesions were carried out within the same patients to account for possible intersubject and interscan differences.
Statistical Analysis
Considering the exploratory nature of the study and the small sample size, no formal statistical analyses were conducted. Descriptive statistics are provided.
Results
The baseline characteristics of the patients included in this study are presented in Table 2 . The average interval between visits was 7.6 weeks, and the average total follow‐up period was 36 weeks.
TABLE 2 Patient characteristics. Patient Sex Age Disease subtype Time since disease onset (months) Number of lesions Number of new lesions during the follow‐up 1 M 23 CIS 5 56 9 2 F 39 RRMS 261 22 1 3 F 35 CIS 1 32 2 4 F 27 RRMS 25 49 8 Abbreviations: CIS, clinically isolated syndrome; F, female; M, male; RRMS, relapsing‐remitting multiple sclerosis. John Wiley & Sons, Ltd.
Patient characteristics.
Abbreviations: CIS, clinically isolated syndrome; F, female; M, male; RRMS, relapsing‐remitting multiple sclerosis.
Cross‐Sectional Analysis
In total, 159 pre‐existing white matter lesions were identified on the baseline scans. A patient‐wise breakdown of all rim‐positive lesions is provided in Table 3 . Examples of T 1 ‐dark and QSM rims are shown in Figure 1 .
TABLE 3 A summary of all rim‐positive lesions recorded at baseline and follow‐up. Number of rim‐positive lesions Baseline Final follow‐up (36 weeks later) Patient T 1 ‐dark QSM T 1 ‐dark QSM 1 19 8 14 7 2 1 0 1 0 3 2 2 2 2 4 0 1 2 2 Total lesions (%) 22 (13.8%) 11 (6.9%) 19 (11.9%) 11 (6.9%) Abbreviation: QSM, quantitative susceptibility mapping. John Wiley & Sons, Ltd.
A summary of all rim‐positive lesions recorded at baseline and follow‐up.
Abbreviation: QSM, quantitative susceptibility mapping.
Lesions with T 1 ‐dark rims on magnetization prepared 2 rapid gradient‐echo (marked by orange arrows) (left) and the corresponding quantitative susceptibility mapping (QSM) image showing a single paramagnetic rim lesion (marked by purple arrow) with a clear vein passing through its center (right). A magnification of the QSM‐rim is provided in the top left corner.
T 1 ‐Dark Rims Are More Common Than QSM Rims
A total of 22 lesions (13.8%) were classed as having a T 1 ‐dark rim and 11 (6.9%) as having a QSM rim. A further six lesions had a clear T 1 ‐dark rim on one slice only and so were classified as rimless according to our protocol.
A total of 17 lesions had a T 1 ‐dark rim only, six had a QSM rim only, and five lesions had both types of rims. Compared to the signal intensity of nearby white matter lesions, five of the six lesions with a QSM rim only appeared globally very hypointense on the MP2RAGE scans and one had patches of hypointense signal.
Lesions With Rims Are Larger Than Rimless Lesions
Figure 2 shows the volumes of all lesions on the baseline scans according to rim status (T 1 ‐dark, QSM, or both). The largest lesions had both T 1 ‐dark and QSM rims, while most of the rimless lesions were noted to have a volume under 100 mm 3 .
Baseline (pre‐existing) lesions with a T 1 ‐dark rim only ( n = 17), quantitative susceptibility mapping (QSM) rim only ( n = 6), or both T 1 ‐dark and QSM rims ( n = 5) had a larger fluid‐attenuated inversion recovery (FLAIR) volume than lesions without either rims ( n = 131). n , number of lesions.
Longitudinal Analysis of Pre‐Existing Lesions
A comparison of the total and patient‐wise numbers of pre‐existing lesions with T 1 ‐dark and QSM rims over a short‐term follow‐up of 36 weeks is shown in Table 3 .
Lesions With T 1 ‐Dark and QSM Rims Show Dynamic Changes on Short‐Term Follow‐Up
On the final follow‐up scan, a total of five pre‐existing lesions had both a T 1 ‐dark and QSM rim. Two lesions had T 1 ‐dark and QSM rims persisting from baseline to follow‐up. Three other lesions had a T 1 ‐dark rim at baseline, but the QSM rim only appeared on the follow‐up scan. Four other T 1 ‐dark rims developed during the study without QSM rims. At baseline, one of these appeared markedly hypointense on the MP2RAGE scan. Table 4 summarizes the temporal evolution of all pre‐existing lesions with a T 1‐ dark or QSM rim at the first and/or last study visit. Out of all the pre‐existing lesions studied, four were possible PRLs and 11 were true PRLs.
TABLE 4 Summary of the longitudinal evolution of lesions with T 1 ‐dark rims and paramagnetic rim lesions between baseline and last follow‐up. Rim lesion type Number of rims persisting Number of rims appearing Number of rims disappearing T 1 ‐dark 14 5 8 QSM 7 4 4 Abbreviation: QSM, quantitative susceptibility mapping. John Wiley & Sons, Ltd.
Summary of the longitudinal evolution of lesions with T 1 ‐dark rims and paramagnetic rim lesions between baseline and last follow‐up.
Abbreviation: QSM, quantitative susceptibility mapping.
T 1 ‐Dark Rim Lesions Can Precede the Appearance of QSM Rims
Four pre‐existing lesions had a QSM rim appear at the final visit. Three of those QSM rims (75%) were preceded by a T 1 ‐dark rim at baseline. The fourth lesion presented patches of hypointense signal in the lesion core on MP2RAGE at baseline. One baseline lesion with a QSM rim had a marked T 1 hypointense signal throughout the lesion core and developed a T 1 ‐dark rim at the follow‐up visit.
T 1 ‐Dark Rims Persist Longer Than QSM Rims
In two of the five (40%) pre‐existing lesions with both a T 1 ‐dark rim and a QSM rim at baseline, the rims persisted 36 weeks later. For the remaining three lesions (60%), the T 1 ‐dark rim persisted for the whole study period, but the QSM rim disappeared.
Longitudinal Analysis of New Lesions
A total of 20 new lesions appeared over the course of the study. Figure 3 describes the T 1 ‐dark and QSM rim status of each new lesion.
The temporal evolution of new lesions with T 1 ‐dark and quantitative susceptibility mapping (QSM) rims. The lesions are arranged according to volume (from smallest to largest). Largest lesions had both T 1 ‐dark and QSM rims, while smaller lesions had a transient T 1 ‐dark rim or were markedly hypointense on the magnetization prepared 2 rapid gradient‐echo image. No new lesions had a QSM rim without also having a T 1 ‐dark rim. Lesions not analyzable on QSM across the whole study duration were excluded. Two lesions were not analyzable at Visit 5 due to poor image quality. Lesions without data for visit 6 are all from Patient 4, who only had five visits.
Most New Lesions Have a T 1 ‐Dark Rim
Eleven lesions met the criteria for a T 1 ‐dark rim, while a further five had a T 1 ‐dark rim on one slice only and were classified as rimless. In eight of the 11 nascent lesions with a T 1 ‐dark rim, the rim appeared at the first appearance of the lesion (see Figure 3 ). Five of the T 1 ‐dark rims persisted for at least three visits and were present at the end of the data collection period. Another four T 1 ‐dark rims disappeared after a single visit, while two lesions with T 1 ‐dark rims only appeared on the final visit and consequently were not able to be followed up.
New PRLs Are Accompanied by a T 1 ‐Dark Rim
Five new lesions (25%) had a rim on QSM (see Figure 3 ). Four of these QSM rims appeared simultaneously to a T 1 ‐dark rim, and the fifth QSM rim appeared one visit after the appearance of a T 1 ‐dark rim. No new lesions had a QSM rim without also having a T 1 ‐dark rim. Of the five new lesions with QSM and T 1 rims, four had both rims still present at the final visit, while one lesion showed the QSM rim disappearing at the final visit with the T 1 rim still present. Two of the five lesions with QSM rims had a rim visible across three visits and were therefore classed as true PRLs.
The Presence of Rims Is Associated With the Size of the New Lesion
The presence of rims was related to the size of the lesions, as measured on FLAIR and MP2RAGE scans (see Figure 4 ). The largest lesions had both T 1 ‐dark and QSM rims (see Figure 4A,C ). Medium‐sized lesions had T 1 ‐dark rims only. The smallest lesions were rimless. The persistence of T 1 ‐dark rims was also related to the size of the lesion (see Figure 4B,D ).
Magnetization prepared 2 rapid gradient‐echo (MP2RAGE) (top row) and fluid‐attenuated inversion recovery (FLAIR) (bottom row) volumes of new lesions according to rim status. (A) Average MP2RAGE volume of new rimless, T 1 ‐dark rim only, and T 1 ‐dark and quantitative susceptibility mapping (QSM) rim lesions. Three lesions without a T 1 rim were unable to be assessed on QSM and were therefore excluded. (B) Average MP2RAGE volume of new lesions according to the temporal evolution of the T 1 ‐dark rim: rimless, temporary T 1 ‐dark rim, or persistent T 1 ‐dark rim. Two lesions with rim appearing at the final visit were excluded. (C) Average FLAIR volume of new rimless, T 1 ‐dark rim only, and T 1 ‐dark and QSM rim lesions. (D) Average FLAIR volume of new lesions according to the temporal evolution of the T 1 ‐dark rim: rimless, temporary T 1 ‐dark rim, or persistent T 1 ‐dark rim.
MP2RAGE and B1 Maps Show High Agreement in Detecting T 1 ‐Dark Rims
T 1 ‐dark rims had consistently longer T 1 values than the lesion core (data not shown). New lesions on B 1 maps corresponded with what was observed on MP2RAGE in 18 of the 20 lesions (90%). One lesion only showed a rim on one slice on the MP2RAGE (and hence had been classed as rimless) but showed a rim on two slices on the B 1 map. The other discrepancy was a large edematous lesion, which was classed as having a T 1 ‐dark rim on MP2RAGE but was globally hyperintense on B 1 .
The Core of New Rimless Lesions Is as Hypointense as the Rim of T 1 ‐Dark Rim Lesions
All the new lesions without a T 1 ‐dark rim were small with a volume of less than 100 mm 3 (see Figure 4 ). On MP2RAGE images, their core appeared diffusely hypointense with signal intensity similar to the rim rather than the core of the large, T 1 ‐dark rim lesions. Calculations of T 1 signal from B 1 maps confirmed these findings. Moreover, new rimless lesions appeared to have higher T 1 values when compared to similar‐sized, pre‐existing, rimless lesions. Example values from Patient 1, who had at least three lesions of each type, are shown in Figure 5 . Consistent with pre‐existing small lesions having reduced T 1 values, the T 1 values of the small new lesions decreased over time (data not shown).
T 1 values of new rimless lesions (whole lesion), lesion core and rim of new T 1 ‐dark rim lesions (provided separately), and pre‐existing rimless lesions (whole lesion) for Patient 1. X shows mean values.
Discussion
This 7T MRI study presents novel findings on the prevalence and short‐term evolution of a newly described imaging marker, T 1 ‐dark rims, and their relationship with PRLs detected on QSM. Our findings suggest that T 1 ‐dark rims are sensitive at detecting, what we believe are, active MS lesions (either acute or chronic active).
We confirm observations from the first study to describe T 1 ‐dark rims that they are more numerous than PRLs [ 11 ]. Using 7T imaging and QSM, we detected twice as many T 1 ‐dark rims compared to PRLs, whereas in the 3T Naval‐Baudin et al. clinical study, T 1 ‐dark rim lesions outnumbered PRLs by around a quarter. In addition to the MRI field strength, our increased T 1 ‐dark rim detection rate is likely due to the differences in our methodologies; Naval‐Baudin and colleagues used SWI and phase imaging, which are not as pathologically specific to iron and, particularly phase, can lead to false positive findings [ 10 ]. In addition, by analyzing all brain lesions in a few patients with early MS, we were able to report the previously unknown prevalence of T 1 ‐dark rims (14% of all lesions).
Our investigation of new lesions revealed a striking prevalence of T 1 ‐dark rims. Over half of the new lesions displayed a T 1 ‐dark rim, a proportion that rose to 80% when considering lesions exhibiting a rim on at least one slice. Our quantitative analyses confirmed visual observations that T 1 ‐dark rims have longer T 1 values than their lesion core. Several pathological processes associated with new lesion development can increase T 1 signal, including demyelination, edema, phagocyte invasion, and hypercellularity [ 4 , 18 ]. Based on the location of the signal, at the periphery of new lesions, we postulate that marked edema associated with acute inflammation, rather than demyelination, is the primary driver of the T 1 ‐dark rim signal [ 19 ].
Notably, the remaining new lesions without a T 1 ‐dark rim had small volumes, all under 100 mm 3 . Small lesions may lack visible rims due to their size (partial volume artifacts), and this could be tested by imaging at higher spatial resolution. However, comparing the T 1 values of these small, new, rimless lesions to the rims and cores of larger T 1 ‐dark rim‐positive lesions confirmed that their core T 1 values matched the values observed at the T 1 ‐dark rim, suggesting a shared pathological origin. Furthermore, small, new, rimless lesions had higher T 1 values than size‐matched, rimless, older lesions, possibly indicating decreased inflammation and edema, consistent with pathological findings of lesions in the chronic inactive stage [ 20 ]. A subsequent T 1 reduction over the follow‐up period likely reflects this progression.
T 1 ‐dark rims were either transient (visible at only one visit) or persistent. A quarter of new lesions had transient T 1 ‐dark rims, suggestive of subsiding edema. These lesions never developed a paramagnetic rim, suggesting a less aggressive disease pathology. Persistent T 1 ‐dark rim lesions were always larger and accompanied by a paramagnetic rim, appearing either concurrently or one visit later. This indicates the presence of ongoing inflammation and tissue destruction, suggesting T 1 ‐weighted imaging may offer a sensitive alternative for CAL detection, particularly as SWI is not currently recommended for MS monitoring.
While paramagnetic rims reflect iron‐laden macrophages and microglia, iron decreases T 1 signal [ 21 ], opposing the increased T 1 signal observed in T 1 ‐dark rims. Thus, the T 1 ‐dark rim is visible despite, not because of, iron accumulation. We speculate that paramagnetic rims may mark a more intense inflammatory and destructive process involving demyelination and macrophage‐mediated myelin (and iron) clearance, making them less common than inflammation alone at acute lesion borders. In another paper by Naval‐Baudin et al., the authors showed that deeply hypointense T 1 voxels, often in the lesion periphery, were almost exclusively present in PRLs compared to non‐PRLs [ 22 ]. The association of both T 1 ‐dark and paramagnetic rims with larger lesion sizes supports the hypothesis that both rim types are associated with a more aggressive disease pathology. These lesions may also represent a subset of slowly expanding lesions, a possibility requiring further investigation with longer follow‐up studies.
Future pathology and MRI studies will establish the role and long‐term evolution of T 1 ‐dark rims in MS. For now, it is interesting to speculate whether the high incidence of T 1 ‐dark rims in new lesions could potentially be used to provide evidence of dissemination in time (DIT) in patients presenting with suspected MS. Future studies could determine whether the simultaneous presence of rim‐positive and rim‐negative lesions can be used as evidence of DIT, similarly to the way gadolinium‐enhancing lesions are currently used. Moreover, T 1 ‐dark rims could have a role in the monitoring of lesions or disease progression. Particularly, T 1 ‐dark rims could act as surrogate markers of early and chronic inflammation and could therefore be used to assess the efficacy of novel disease‐modifying therapies such as Bruton's tyrosine kinase inhibitors, which act on both sides of the BBB. Given the existing, widespread use of T 1 ‐weighted imaging in routine MS monitoring and the detection of T 1 ‐dark rims by clinical 3T scanners [ 11 ], this new radiological sign could easily be implemented in clinical practice.
This analysis faces several limitations. First and foremost, our exploratory study included only four patients, and our results need to be validated in a larger sample size. Importantly, these data did not include contrast‐enhanced imaging, so it was not possible to definitively identify the acute stage of nascent lesions or determine a temporal correlation between T 1 ‐dark rims and the resolution of gadolinium enhancement. However, it is known that MS lesions rarely enhance beyond 2–3 months [ 23 ], and T 1 ‐dark rims were seen to persist for 5 months, until the end of the data collection period.
Furthermore, since nascent lesions with a persistent T 1 ‐dark rim and paramagnetic rim were still observed at the end of the study, interpretation of the long‐term evolution of T 1 ‐dark rims and the temporal correlation with paramagnetic rims is limited. Further longitudinal follow‐up and inclusion of imaging techniques such as proton density, to confirm whether T 1 ‐dark rims represent edema, and diffusion‐weighted imaging, to infer the underlying tissue integrity, will be useful in discerning the underlying pathology of T 1 ‐dark rims and differences between them and PRLs.
This study further characterized the novel imaging finding of T 1 ‐dark rims that surround some MS lesions. We propose that T 1 ‐dark rims may represent a marker of early lesions with ongoing edema/inflammation at the edges surrounding demyelination. Considering that T 1 ‐weighted imaging is one of the most commonly utilized sequences, T 1 ‐dark rims could easily be incorporated into routine, clinical assessment of patients with MS. However, future large, longitudinal studies are needed to further establish their role, evolution, and underlying pathology.
Conflicts of Interest
The authors declare no conflicts of interest